manage-latex-bibliography
Automates bibliographic management for LaTeX projects, including auditing and scanning citations.
Install
mkdir -p .claude/skills/manage-latex-bibliography && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/11622" && unzip -o skill.zip -d .claude/skills/manage-latex-bibliography && rm skill.zipInstalls to .claude/skills/manage-latex-bibliography
Activation
This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.
Use when a LaTeX project needs a new or updated BibTeX bibliography, missing citation entries, verified metadata, headline-style titles, or optional AEA bibliography-style setup.Key capabilities
- →Verify every .bib entry online for initialization.
- →Check .tex files against .bib for missing keys.
- →Audit .bib files against a references directory for missing PDFs.
- →Force update a specific BibTeX entry.
- →Cross-check existing .bib entries against Crossref for metadata drift.
- →Apply approved changes to the bibliography.
How it works
The skill performs various checks on LaTeX bibliography files, verifies metadata against online sources, and applies approved updates.
Inputs & outputs
When to use manage-latex-bibliography
- →Fixing missing BibTeX keys
- →Verifying publication metadata
- →Auditing bibliography vs PDF directory
- →Initializing new project bibliographies
About this skill
Manage LaTeX Bibliography
Workflow
-
Infer the operation from the request and available files:
- Scan checks
.texfiles against the bibliography for missing keys. - Audit checks the bibliography against the reference PDF directory.
- Initialize runs
audit --allto verify every bibliography entry; use it only when the user explicitly requests initial or full verification. - Verify existing checks existing entries against Crossref for DOI, title, and year drift.
To revise a single entry, edit its object in the generated proposal JSON directly, then re-run
validate.When the request clearly identifies one operation, proceed without asking. Ask one concise clarification only when the operation or required scope cannot be inferred.
- Scan checks
-
Confirm only the inputs required for that operation: the project root, main
.texfile,references.bib, reference PDF directory, or proposal file. -
Locate this skill directory and assign its absolute path to
SKILL_DIR. -
Run the inferred operation without modifying the project:
Scan:
python "$SKILL_DIR/scripts/manage_bibliography.py" scan \ --project /path/to/project \ --output /path/to/project/bibliography-proposal.jsonInitialize:
python "$SKILL_DIR/scripts/manage_bibliography.py" audit \ --bib /path/to/project/references.bib \ --pdf-dir /path/to/project/references \ --output /path/to/project/bibliography-proposal.json \ --allAudit:
python "$SKILL_DIR/scripts/manage_bibliography.py" audit \ --bib /path/to/project/references.bib \ --pdf-dir /path/to/project/references \ --output /path/to/project/bibliography-proposal.jsonVerify existing:
python "$SKILL_DIR/scripts/manage_bibliography.py" verify-existing \ --bib /path/to/project/references.bib \ --output /path/to/project/bibliography-verify-report.jsonThe report lists each entry with
verified,inconsistent,unverified, orskippedstatus. If Crossref is unreachable, entries gracefully degrade tounverifiedso the workflow does not crash. -
Review missing citation keys and inspect prose for likely uncited works. Never treat a citation key or prose mention as proof of publication identity.
-
Research each candidate according to
references/verification-rules.md. Use any available skill that provides theliterature-searchcapability (e.g., literature-search-repec, literature-search-openalex) to retrieve the ground-truth metadata. If no such skill is installed, fall back to theverify-existingCrossref path or tag the entry[UNVERIFIED]and notify the user rather than fabricating metadata. Seedocs/architecture-dependencies.mdfor the graceful-degradation contract. -
Give each candidate and its evidence to an independent subagent for a fresh online check. If subagents are unavailable, perform a separate second lookup without relying on first-pass conclusions.
-
Format titles according to
references/title-case-rules.md. -
Populate each proposal entry with its BibTeX type, fields, source URLs, conflicts, verifier identity, and status. Use
verifiedonly after the independent check succeeds. -
Ask the user to approve every prose-inferred reference and every correction to an existing entry. Corrections must include the existing fields as
before_fields, useapprovedstatus, and set both approval flags to true. -
Validate the completed proposal:
python "$SKILL_DIR/scripts/manage_bibliography.py" validate \ --proposal /path/to/project/bibliography-proposal.json -
Apply only after validation succeeds:
python "$SKILL_DIR/scripts/manage_bibliography.py" apply \ --proposal /path/to/project/bibliography-proposal.json -
Review
bibliography-apply-result.jsonand compile the project when a TeX toolchain is available.
Re-run scan whenever a tracked .tex or .bib file changes.
Entry Rules
Use lowercase authorYearFirstTitleWord citation keys, for example
acemoglu2001colonial. Skip title stop words. If the key collides with another
work, append the next meaningful title word; do not add an arbitrary numeric
suffix.
Use lowercase BibTeX field names and string field values. Keep braces balanced and protect case-sensitive title content before validation.
Inferred References
The agent may identify author-year mentions, titles, DOI strings, or other clear publication references in prose. Show the source passage and proposed publication to the user. Do not add the citation or entry until the user approves it.
AEA Style
This skill does not distribute aea.bst. For a traditional BibTeX project,
ask the user before downloading the official AEA template. After confirmation:
python "$SKILL_DIR/scripts/manage_bibliography.py" install-aea-style \
--project /path/to/project \
--confirm-download
The helper accepts only an HTTPS download that ends on the official
aeaweb.org host and records its URL, retrieval time, and SHA-256 digest.
Never use an unofficial mirror.
Preserve an existing bibliography style. For biblatex or biber, maintain
entries but do not activate aea.bst or convert the project.
Safety Rules
- Never invent bibliographic metadata.
- Never apply unresolved or unverified entries.
- Never apply inferred references or existing-entry corrections without user approval.
- Do not bypass duplicate identifier, stale digest, project containment, archive, or overwrite failures.
- Do not add
aea.bstor downloaded AEA files to this skill or repository.
When not to use it
- →When the user wants to invent bibliographic metadata.
- →When the user wants to apply unresolved or unverified entries.
- →When the user wants to apply inferred references or existing-entry corrections without approval.
Prerequisites
Limitations
- →Never invent bibliographic metadata.
- →Never apply unresolved or unverified entries.
- →Never apply inferred references or existing-entry corrections without user approval.
How it compares
This skill provides a structured workflow for auditing, verifying, and updating LaTeX bibliographies, ensuring data integrity and consistency, unlike manual bibliography management.
Compared to similar skills
manage-latex-bibliography side by side with the closest alternatives in the catalog.
| Skill | Installs | Updated | Safety | Difficulty |
|---|---|---|---|---|
| manage-latex-bibliography (this skill) | 0 | 2mo | Review | Intermediate |
| ml-paper-writing | 48 | 6mo | Review | Advanced |
| content-research-writer | 15 | 10mo | No flags | Beginner |
| research-grants | 6 | 7mo | Review | Advanced |
Try saying
Example prompts that trigger this skill in your AI assistant.
You might also like
ml-paper-writing
davila7
Write publication-ready ML/AI papers for NeurIPS, ICML, ICLR, ACL, AAAI, COLM. Use when drafting papers from research repos, structuring arguments, verifying citations, or preparing camera-ready submissions. Includes LaTeX templates, reviewer guidelines, and citation verification workflows.
content-research-writer
ComposioHQ
Assists in writing high-quality content by conducting research, adding citations, improving hooks, iterating on outlines, and providing real-time feedback on each section. Transforms your writing process from solo effort to collaborative partnership.
research-grants
davila7
Write competitive research proposals for NSF, NIH, DOE, and DARPA. Agency-specific formatting, review criteria, budget preparation, broader impacts, significance statements, innovation narratives, and compliance with submission requirements.
nsfc-research-content-writer
huangwb8
为 NSFC 正文"(二)研究内容"写作/重构,并同步编排"特色与创新"和"三年年度研究计划",输出可直接落到 LaTeX 模板的三个 extraTex 文件。适用于用户要写/改"研究内容、研究目标、关键科学问题、技术路线、创新点、三年计划/里程碑"等场景。
clinical-reports
davila7
Write comprehensive clinical reports including case reports (CARE guidelines), diagnostic reports (radiology/pathology/lab), clinical trial reports (ICH-E3, SAE, CSR), and patient documentation (SOAP, H&P, discharge summaries). Full support with templates, regulatory compliance (HIPAA, FDA, ICH-GCP), and validation tools.
business-knowledge-workflow
TencentBlueKing
业务知识获取与 Skill 文档编写工作流。当用户需要熟悉新业务模块、从 iWiki 获取文档、结合代码分析生成架构文档、或将业务知识沉淀为 Skill 时使用。